Keywords

1 Introduction

Lowland peatlands are a unique ecosystem in Southeast Asia, distributed mostly in the east coast of Sumatra, Central and West Kalimantan, Sarawak, and southern Papua (Hooijer et al. 2010). Peat soil consists mainly of decomposed trunks, branches and roots of trees which have been sedimented over a period of 4000–26,000 years (Page et al. 1999, 2004). Peatlands are nutrient poor and have extremely acidic soil (pH 3.0–4.5) because they are situated away from large rivers that bring mineral nutrients downstream from mountainous areas. The total area of peatlands in Southeast Asia is estimated to be 24.8 million ha, comprising 56% of the global total. Southeast Asian peatlands are estimated to store 68.5 Gt of carbon, comprising 11–14% of the carbon stock in the global peatland (Page et al. 2011).

The natural vegetation of Southeast Asian peatlands is peat swamp forest (Anderson 1961; Bruenig 1990; Posa et al. 2011). Peat swamp forests in Southeast Asia function not only as a significant carbon sink, but also as a refuge for many endangered species in the region. The faunal composition of this vast ecosystem has not been as well studied as the floral composition, however (Gaither Jr 1994; Page et al. 1997; Whitemore 1984). Because of the poor nutrient content of the peat soil and the low primary productivity of the forests, they have been presumed to have lower diversity and abundance of animals than surrounding lowland forest (Janzen 1974; Posa et al. 2011; Whitten et al. 2000). Many endangered species, however, are known to inhabit peat swamp forests (Buckley et al. 2006; Cheyne et al. 2008; Felton et al. 2003; Morrogh-Bernard et al. 2003; Yule 2010). Furthermore, Ehlers Smith and Ehlers Smith (2013), Johnson et al. (2005), and Quinten et al. (2010), indicated that densities of several primate species, including orangutan, in peat swamp forests were higher than those in adjacent lowland forests. Gaither Jr (1994) also detected that some bird species were more abundant in peat swamp forest than in lowland forest.

Recently, ground-dwelling mammals and birds in natural peat swamp forest have been investigated with camera traps. Camera traps are automatic digital cameras equipped with infrared motion sensors that present new methods for investigation of faunal diversity (Cannon et al. 2007; Cheyne et al. 2010; Cheyne and Macdonald 2011; Posa 2011; Kays and Slauson 2008; Rowcliffe and Carbone 2008; McCallum 2012; Burton et al. 2015; Nakashima et al. 2018). Mohd-Azlan (2004) conducted camera-trapping studies in a natural peat swamp forest at Maludam National Park in Sarawak; Cheyne et al. (2010) and Cheyne and Macdonald (2011) conducted the same in the Sebangau National Park in Central Kalimantan; and Posa (2011) along the Kapuas River, Central Kalimantan. However, some of those studies reported only felid and bird species, and the others ran their camera traps for only a limited number of days.

In recent decades, large-scale plantations of fast-growing trees and oil palm have been developed in the peatlands in Southeast Asia (Uryu et al. 2008; Corlett 2009; Yule 2010). Posa et al. (2011) estimated that a minimum of 64% of the historical peat swamp forest in Southeast Asia has been lost so far. The deforestation and drainage of the peatlands results in huge emission of carbon dioxide, which accelerates global warming (Couwenberg et al. 2009). Carbon dioxide emissions from peat swamp forest degradation in Southeast Asia during 1997–2006 was estimated to be as much as 0.30 Pg C year−1, or 3% of total global anthropogenic carbon emissions (van der Werf et al. 2009).

Conversion of natural peat swamp forests to industrial tree plantations and oil palm plantations can also cause significant loss of biodiversity, but this phenomenon is not yet well studied (Posa 2011; Posa et al. 2011). Some animal species inhabiting peat swamp forests may be able to adapt to the new vegetation (Meijaard et al. 2010), but plantantions may also lead to serious loss of biodiversity, especially among species vulnerable to development in surrounding lowland forests in the region (Fitzherbert et al. 2008; McShea et al. 2009). Understanding the effects of plantation development on the animal community in peat swamp forests is necessary for proper management of the peat swamp landscape of the region.

To understand the faunal diversity of natural peat swamp forests and the impact of plantation development on these, we investigated species composition and abundance of ground-dwelling, medium to large mammals and birds using camera traps in natural and planted acacia forests in Giam Siak Kecil-Bukit Batu Biosphere Reserve (GSK-BB), Riau Province, Indonesia, in the Island of Sumatra. Some of the results were already published as Samejima et al. (2016).

2 Materials and Methods

2.1 Study Area

The Giam Siak Kecil-Bukit Batu Biosphere Reserve (GSK-BB) is located along the Strait of Malacca, covering an area of 7053 km2. This biosphere reserve was established by a private-sector initiative and was registered by UNESCO’s Man and Biosphere (MAB) program in 2009. GSK-BB consists of three zones: a core area (1787 km2), buffer zone (2224 km2), and transition area (3041 km2). The core area consists natural peat swamp forest mostly in the Giam Siak Kecil Wildlife Reserve and the Bukit Batu Wildlife Reserve. The buffer zone comprises planted forests of fast-growing tree species (Acacia crassicarpa and A. mangium) inside four industrial tree plantations, called Hutan Tanaman Industri (HTI) in Indonesia. The plantations also have natural forests in areas adjacent to the Wildlife Reserves, which constitute a part of the core area. The natural forest was preserved in accordance of the Ministry of Forestry Decree No.70/Kpts-II/1995 and No.246/Kpts-II/1996, requiring that 10% of a concession area should be protected as “Kawasan Lindung.” The outermost transition area of GSK-BB contains agricultural fields of smallholders and some oil palm plantations.

This study was conducted in the Bukit Batu area (BB) in the northeastern part of GSK-BB covering an area of approximately 800 km2 (Fig. 4.1, Fujita et al. 2016). BB is located at latitude 1° 17–28′ north and longitude 101° 42′–102° 00′ east, including Bukit Batu Wildlife Reserve and industrial tree plantations of Bukit Batu Hutan Alam (PT. BBHA) and Sekato Pratama Makmur (PT. SPM). All of the study area comprises 4–8 meter-thick peat (Simbolon et al. 2011), with a water level of about 40–100 cm below ground (Kozan 2016).

Fig. 4.1
The layout of the Bukit Batu area includes camera set points, open areas, natural peat swamp forest, secondary forest, planted acacia forest, protected areas, village, roads, and river.

Location of ten plots inside Bukit Batu area. Five random points were selected in each plot to set camera traps

Full size image

The annual amount of rainfall in the study area is highly variable and its seasonality was unclear. Annual rainfall at Pekanbaru, located 100 km southwest of BB, ranged from 102 mm to 284 mm from 1999 to 2001, but was 1074–2596 mm from 2003 to 2011 (National Climatic Data Center 2012). Clear dry periods existed in some years, but the timing of such dry periods varied every year. Temperature ranged from 23 °C to 32 °C throughout the year.

The core zone in our study area was mostly covered by natural peat swamp forest. However, large trees within a few kilometers of both banks along the lower part of the Bukit Batu River in the Wildlife Reserve were logged by a logging company and smallholders during 1990s and 2000s (Watanabe et al. 2016). In the Wildlife Reserve, Gunawan et al. (2012) established two 0.5-hectare plots and measured all trees with diameter at breast height (DBH) > 3 cm. In the protected area, Partomihadjo et al. (2011) established two 1-ha plots and measured all trees with DBH > 10 cm. Tree species composition in these four plots were similar, the common major species being Palaquim sumatrana/dasyphyllum, Syszigium paludosa, Diospyros hermaphroditica, Mezzetia parvifolia, and Madhuca motleyana. The forest was categorized as mixed swamp forest following the classification by Anderson (1976).

The buffer zone comprises planted forest of A. crassicarpa with networks of canals for the water management and transportation constructed at intervals of 500–3000 m. A. crassicarpa is a fast-growing species well adapted to peat soil. It grows in dense stands within 3 years of planting. The two plantations in our study area were established during 1999 (Watanabe et al. 2016) and developed until 2006. Acacia trees are harvested and replanted on a 5–6-year rotation, and these planted acacia forests consist of patches of even-aged trees. The forest floor of the planted acacia forest is dominated by Dicranopteris sp. and Blechnum sp.

Hunting pressure on medium to large mammals in this area is considered to be low because there are no residents inside the study area. Also, most of the nearby inhabitants are Muslim (Suzuki et al. 2016) who rarely hunt wild mammals. Nevertheless, we found snares to catch wild boar set by outsiders and the skull of a sambar deer (Rusa unicolor) at a camp used by poachers in the Wildlife Reserve. Cause of death of the sambar deer—either at the hands of poachers or natural causes—was unclear.

2.2 Camera Trapping

To evaluate the species composition and abundance of ground-dwelling mammals and birds, we deployed camera traps—automatic digital cameras with infrared motion sensors (Bushnell Trophy Cam, Model 119435, Bushnell, Olathe, KS), in BB, from December 2010 to October 2011.

We established four plots in the natural peat swamp forest inside the Bukit Batu Wildlife Reserve (WR1, WR2, WR3, WR4) and three plots in the natural peat swamp forest inside the protected areas of the two industrial tree plantations (PA1, PA2, PA3) (Fig. 4.1 and Table 4.1). While WR4 and PA1–PA3 were in primary forests, WR1–WR3 were in forests degraded by previous logging activities. WR1–WR4 were located about 500–1000 m from the banks of Batu River. The altitudes of WR1–WR3 were lower than those of all the other plots. PA1–PA3 were located away from the river, but located close to canals (approximately 500–1000 m) in the plantations. We also established three plots in the planted forests of A. crassicarpa inside the two plantations (AF1, AF2, AF3). They are all close to canals (approximately 500 m). Each of the ten plots covered a circular area with a radius of 500 m. The altitudes of these plots ranged from 8 m to 40 m above sea level. The forest floor was generally flat and covered in wooden debris. Ground surface of the plots were not flooded during the study period.

Table 4.1 Longitude and latitude of centers of the ten plots
Full size table

We set camera traps at five random points in each plot and compared the species compositions captured at each plot and a mean trapping rate (MTR) for each species. The camera was mounted onto a tree at a height of 50–100 cm above ground. We directed the camera to focus downward to capture a field-of-view of approximately 2–7 m2. We set the camera to record in video mode for a duration of 10 s upon trigger. Once triggered, a camera could not be triggered again for 10 s. The camera was inspected every 3–5 months to change batteries and memory card.

All medium to large animals including mammals, terrestrial birds, and monitor lizard species captured by the camera traps were recorded and identified to species level based on Francis (2008), MacKinnon and Phillipps (1993), and Payne and Francis (2005). We classified wild boar as a morph-species (Sus sp.) because clear distinctions could not be made between Sus scrofa and S. barbatus, and some resembled a crossbreed between the two species. We defined a record as an independent record if the same species was not recorded during the previous 30 min. We counted the number of independent records of each species and the total length of camera-working days at each set point.

Conventionally, the trapping rate is the total number of independent records per total number of camera-working days multiplied by 100, and is termed as Relative Abundance Index (RAI) by O’Brien et al. (2003). As explained by Rowcliffe et al. (2008, 2013) and Nakashima et al. (2018), RAI can provide a index of animal density if the captured areas of a camera trap, animal speeds, and lengths of animal activity time in a day on average are not different among the plots. However, this value can be biased by the difference of camera-working days among the set points inside a plot. To offset the bias, we calculated a mean trapping rate (MTR) as an index of animal density as follows: a single-liner regression equation between the length of camera-working days at each set point and the number of independent records, with the regression line fixed to cross the point of origin. We multiplied the coefficient by 100 and defined the value as the MTR of the plot.

2.3 Effects of Planted Acacia Plantations

We evaluated the effects of vegetation differences caused by conversion to planted acacia forest on the trapping rate of each species in BB using the model selection method. We also compared the similarity of species composition between plots in different vegetation types using nonmetric dimensional scaling (nMDS).

To evaluate the effect of vegetation differences, we made two nested mixed Poisson models (mixed with a fixed effect and a random effect) for a number of records or number of species captured at each camera set point. The equations of the two models are as follows:

$$ \left(\mathrm{Model}\ 1\right)\ \mathrm{N}=\mathrm{wd}+\mathrm{P}+\mathrm{V} $$
$$ \left(\mathrm{Model}\ 2\right)\ \mathrm{N}=\mathrm{wd}+\mathrm{P}, $$

where ‘N’ denotes the number of records, ‘wd’ is the length of the camera-working days, ‘P’ and ‘V’ are categorical data, representing the Plot ID and the vegetation type of the plot respectively. We compared goodness-of-fit of Model 1 and Model 2 based on Akaike’s Information Criterion (AIC) values (Akaike 1973). AIC is a measure of goodness-of-fit with an added penalty for model complexity as measured by the number of fit parameters in the model (Burnham and Anderson 2002). We selected model with vegetation (model 1) as the best model when AIC of Model 1 was smaller than AIC of Model 2 and the difference was more than two.

On the other hand, nMDS is a robust unconstrained ordination method suited to community data that searches iteratively for best configuration of the plots from the dissimilarities (Gotelli and Ellison 2004). We plotted the ten plots following Bray-Curtis dissimilarity of species composition (the record numbers of each species) to investigate effect of the vegetation differences.

The generation of random set points for the cameras, and all statistical analysis were conducted using the statistical software R 3.0.1. (R Development Core Team 2013) with the packages of spatstat, lme4, and vegan.

3 Results

3.1 Faunal Composition in the Peat Swamp Forest of BB

Total camera-working days were 3978 days at 20 camera setting points in the four plots in the natural forest inside the Wildlife Reserves; 3336 days at 15 points in the three plots in the natural forest inside the protected areas of the HTIs; and 3675 days at 15 points in the three plots in the planted acacia forest inside the HTIs.

During the 10,989 camera-working days at 50 points in total, we obtained 1856 records of 19 species of mammals, 3 terrestrial birds, and 1 monitor lizard (Table 4.2). These included 11 vulnerable or endangered species on the IUCN Red List (IUCN 2012), such as sun bear, Sunda clouded leopard (Neofelis diardi), marbled cat (Paradofelis marmorata), and Sunda pangolin (Manis javanica). The images of all captured species were presented in Fujita et al. (2012).

Table 4.2 Mean trapping rate of each species
Full size table

In addition to the species we recorded by camera traps, we were informed by local people that sambar deer and Sumatran tiger (Panthera tigris) also inhabited the natural peat swamp forest in this area. We found the skull of a male sambar deer at an illegal bird-hunting camp in the core area. It was also reported that a resident was attacked and killed by a tiger in the transition area.

The sum of MTRs of all species ranged from 9.22–51.85 (mean: 29.16), 8.75–31.76 (mean: 16.42), and 2.29–6.38 (mean: 4.02) in the natural forest inside the Wildlife Reserves, the natural forest inside the protected areas of HTIs, and the planted acacia forest inside the HTIs respectively. Camera-trap images were dominated by three species (Table 4.2), namely wild boar (Sus sp.), southern pig-tailed macaques, and lesser mouse-deer (Tragulus kanchil). Wild boar contributed to 43.2%, 69.0%, and 50.7% of all records in the Wildlife Reserves, the protected area of HTI, and the planted acacia forest, respectively. The southern pig-tailed macaques contributed to 16.9%, 19.9%, and 24.5%, respectively. Lesser mouse-deer, which were abundant only in the Wildlife Reserve, represented 33.7% of all the images. All other species had low MTRs compared to these three dominant species.

Among the three terrestrial bird species we detected, crestless fireback (Lophura erythrophthalma) was the dominant species in the natural peat swamp forest, accounting for 82.1% of all the records, while red jungle fowl (Gallus gallus) was recorded only in the planted acacia forests.

3.2 Differences of Faunal Composition Between Natural Peat Swamp Forests and Planted Acacia Forests

The natural peat swamp forests had a higher species richness than the planted acacia forests (Fig. 4.2). The average number of species per plot and the range were 9.75 (9–11) in the Wildlife Reserve, 11 (11) in the protected areas of HTIs, and 5.3 (4–6) in the planted acacia forests inside the HTIs (Table 4.2). Total number of recorded species was 16 in the Wildlife Reserve and 18 in the protected areas of HTIs, and seven in the planted acacia forests inside the HTIs. The higher species richness in natural peat swamp forest was the result of the number of rare species. Among the 11 vulnerable or endangered species recorded in this study, eight were recorded in the natural forests, while the other three were recorded in both vegetation types. Of the 11 vulnerable or endangered species, none were recorded only in the acacia forest.

Fig. 4.2
A graph plots the number of species on the y-axis versus camera trapping days on the x-axis. The 10 lines have an increasing trend.

Species accumulation curve in the ten plots. Solid line; natural forest in the Wildlife Reserve (WR1, WR2, WR3, and WR4 plots). Dash line; natural forest in the protected areas of Hutan Tanaman Industri (HTI) (PA1, PA2, and PA3 plots). Dotted line; planted acacia forest in the HTIs (AF1, AF2, and AF3 plots)

Full size image

The best mixed models selected using AIC for MTRs of 12 species including all of the abundant species in our study area were with the vegetation variable, comprising natural peat swamp forest or acacia forest (Table 4.2). Among the 12 species, the MTRs of nine species were higher in natural forest than that of the planted acacia forest. On the contrary, the MTRs of Malay civet (Viverra tangalunga), leopard cat (Prionailurus bengalensis), and red jungle fowl were higher in planted acacia forest than those in the natural peat swamp forest. Besides, the best mixed model with vegetation variable between the natural peat swamp forests of the Wildlife Reserve and the protected area of HTI, was selected for only three species. Lesser mouse-deer and crestless fireback were more abundant in the Wildlife Reserve, whereas short-tailed mongoose (Herpestes brachyrus) was more abundant in the protected area.

The result of nMDS also showed the similarity of species composition at plots in the Wildlife Reserve and protected areas of HTI, and the dissimilarity from those at plots in the planted acacia forest (Fig. 4.3, stress: 0.084). The four plots in Wildlife Reserve and three plots in protected areas of HTI are overlapping especially on Axis 1, while three plots in the planted acacia forest were plotted apart from them.

Fig. 4.3
A graph plots n M S D 2 on the y-axis versus n M S D 1 on the x-axis. The wildlife reserve, and protected areas overlap each other, while the planted acacia forest is away.

Similarity of species composition among the ten plots shown in nonmetric dimensional scaling

Full size image

4 Discussion

4.1 Ground-Dwelling Mammals and Bird Communities in Peat Swamp Forests

Our results indicated that the natural peat swamp forest in BB is a habitat for various mammal and terrestrial bird species including the vulnerable or endangered species. Notably, wild boar, southern pig-tailed macaques, and lesser mouse-deer were dominant species, accounting for 93% of all records (Table 4.3). One of the authors (HS) also conducted a camera-trapping survey in two lowland forests in East Kalimantan (Ratah Timber and Roda Mas) and one lowland forest in Sarawak (Anap-Muput) (Jati et al. 2018; Samejima and Hon 2019). Species compositions of the four study sites were similar, and wild boar (Sus barbatus), pig-tailed macaques (Macaca nemestrina), and mouse-deer (Tragulus kanchil and T. napu) were the top three or four most frequently recorded species also in the lowland forests (except Tragulus spp. in Anap-Muput). The proportions of wild boar found in the other sites were lower than those found in this study. While species accumulation curves in natural forests in this study were not yet saturated, species composition in peat swamp forest may be simpler than lowland rainforest in same region. Sasidhran et al. (2016) and Adila et al. (2017) also conducted a camera-trap survey in the North Selangor Peat Swamp Forest, Peninsular Malaysia, and recorded 4997 images of medium to large mammals. Among them, wild boar (Sus scrofa) accounted for 55% of all records. These results suggest that the natural peat swamp forest is an ecosystem dominated by wild boar. However, more studies in the natural peat swamp forest in this region are necessary to test these hypotheses.

Table 4.3 Proportion of records of three dominant species in natural forests in this study and three sites of lowland mixed dipterocarp forests in Sarawak and East Kalimantan
Full size table

4.2 Impact of Plantation Development Converted from Natural Peat Swamp Forests on Ground-Dwelling Mammals and Bird Communities

Our results indicate that the development of acacia plantations had a severe impact on species richness and abundances of ground-dwelling mammals and birds inhabiting the natural peat swamp forest. MTRs of most of the species, including the dominant species of wild boar, lesser mouse-deer and southern pig-tailed macaque, were lower in the planted acacia forests than in the natural peat swamp forests, suggesting significant population decrease of these species as a result of land-use conversion from natural peat swamp forests to industrial tree plantations. The obvious difference of species richness and the composition were also observed by Fujita et al. (2016) for bird species in same study area.

The plantation companies created substantial protected areas to maintain original vegetation. Our result showed that faunal diversity in the protected areas was as rich as in the Wildlife Reserve. However, faunal diversity inside the planted acacia forest itself was extremely low. As more than half of the species inhabiting the natural peat swamp forest were not detected in the planted acacia forests, the acacia forests are highly unsuitable as habitat for them. The planted acacia forest cannot function as a corridor for most of the ground-dwelling mammals and birds between the core area and the surrounding natural forests outside of GSK-BB.

McShea et al. (2009) also conducted a camera-trap study in an industrial tree plantation of Acacia mangium located in the lowland areas of central Sarawak in Borneo. In the plantation, many small patches of degraded natural forests were maintained among the planted acacia forests in addition to three large conservation zones. Presence of ground-dwelling mammal species in the remnant natural forests were higher than those in the planted acacia forests in general. Furthermore, mammal presence in the planted acacia forests were positively correlated with the closeness of the camera set points to the natural forests (McShea et al. 2009). The total number of species (n = 20) detected in the planted acacia forests was not much different (n = 24) than in the natural forests. McShea et al. concluded that planted acacia forest may not be an appropriate habitat for most endemic species, many species use the planted acacia forest to transit among patches of remnant natural forests. In short, the mosaic landscape of the acacia-natural forests may enable the persistence of many animal species in the industrial tree plantation. Nasi et al. (2009) also studied primates in an acacia plantation in lowland Sumatra, finding that natural forest fragments within industrial tree plantations do contribute to the maintenance of biodiversity. Further study may make it possible to propose appropriate design of such mosaic landscapes able to support original biodiversity as a feature of plantation development.

Some species such as Malay civet, leopard cat, and red jungle fowl may have adapted to the new plantation environment. McShea et al. (2009) also detected that wild boar (Sus barbatus), civet and mongoose (Herpestes brachyurus, H. semitorquatus, Paradoxurus hemaphroditus and Hemigalus derbyanus), and felidae (Pardofelis marmorata and Prionailurus bengalensis) were found in acacia forests more frequently than in logged natural forest. The adaptability of these species suggests that, whereas frugivorous species may have limited food resources in the planted acacia forest, carnivore species are not much affected.

Planted acacia forests in BB may also function as a buffer zone protecting intact natural forest in the core area from forest fires that frequently occur in the transition zone of BB (Eyes on the Forest 2013). The northern part of the transition zone in BB has burnt repeatedly and become an open area with few scattered-trees (Fig. 4.1). The plantation companies have monitored and prevented the spread of forest fires from the transition zone into their planted acacia forests.

A limited area of natural forest currently remains in lowland Sumatra. Our results indicate that natural peat swamp forest in the core area is vital to maintaining regional biodiversity, and more appropriate landscape designs, like acacia-natural forest mosaic, may boost biodiversity in industrial tree plantations.